History

0.2.0-dev

Dropped support for frontier.frame.DataFrame, electing to return a pandas DataFrame instead.

Introduced frontier.frame.DataFrame which extends numpy's ndarray, that will now be returned from frontier.frontier.Statplexer's API functions that return read in data. * frontier.frame.DataFrame.transform takes a dictionary mapping labels to a lambda function that will be applied to the corresponding column of values in the DataFrame. * frontier.frame.DataFrame.transform allows specification of keyword argument add_unknown=True which will append a zero filled column to the right of the frame. Transformations will then be allowed to continue as if the new column had always existed and can be populated using data from other columns using the transformation syntax. * frontier.frame.DataFrame.exclude allows exclusion of a list of labels, returning a new frame with any applicable columns removed. * frontier.frame.DataFrame.add_observation stacks a new observation array to the DataFrame.
_test_variance function of the Statplexer will now test the range of variance magnitudes across each parameter of read in data, issuing a warning and producing a table if a magnitude difference greater than ±1 is discovered.

Fix #2
- Add get_id to data readers to prevent cluttering parameter space.
- Update TestBamcheckReader.test_id_key to use get_id() instead of get_data()["_id"]

Documentation now exists.
Required data readers to specify an _id instead of forcing use of data file basename in the Statplexer.